I'll be interested to see how you are gathering your info as I know that you're a very precise 'kind of person' and how you interpret it
Well that statement puts the pressure on.
So here is my theory and process. By the way the more I learn about grafana the more I love it.
The concept is really simple divide watts out by watts in and you get efficiency or lack thereof. As in 50 watts out 100 watts in hence 50/100 =0.5 50% efficiency or 90/100 =0.9 or 90% efficient.
How to accomplish this with the sensors and equipment we have at hand.
One piece of equipment we have that is supposed to be "calibrated" is the Batrium Shunt. It in essence measures current flow and voltage. We know we can calculate wattage with the equation V*A=W hence 50V *10A =500 Watts.
The only problem is that the batrium shunt is bidirectional in escence giving us positive and negative readings. Also it is not really a good indicator of actual battery usage as the only time it is giving us an accurate amperage draw is when there is no solar input which is at night.
Additionally at that time it is a negative value and division becomes problematic without going to extensive mathematical formulas.
The other piece of equipment is the IoTaWatt which has CTs on both L1 and L2 giving me the Wattage of my inverters output.
As I subscribe to the kiss principle (well most of the time) I installed a 100A/75mv calibrated shunt that measures the actual amperage draw by the inverters through an ESP32. Just some code and math. If that reading correlates with the negative amp reading of the batrium shunt within a reasonable amount I should be on track , and it does!
So grafana to the rescue.
First things to plot all Watts in and out. Inverters, Grid, Batrium/Batteries, and Solar.
Great OK that gives me a good overall view.
Now the thing we are interested in is basically Inverter watts in and inverter watts out.
Once we get those figures we can then determine efficiency.
To accomplish this I used the grafana transformations feature to do some math with all the sensor input readings.
I came up with this graph which is the information we need watts into the inverters and out of the inverters with positive numbers.
You can see the blue trace (Inverter Output Watts) is just a bit under the red trace (Inverter Input Watts) as expected.
Now it was just a matter of calculation watts out/watts in.
Grafana again to the rescue. It provides relatively instantaneous varying results of 90.64% at this moment and with a graph over time we can get an average which in my case is 90.1937% efficiency from my inverters.
This is very encouraging as the GTIL2 inverters I have claim to have a Peak Inverter Efficiency of 90%.
Kudos for a chinese company actually telling the truth.
And another Calculation.
Now the next challenge is to put Solar Watts into the equation to determine the Theory of Everything which would unify all the fundamental interactions of nature: gravitation, the strong interaction, the weak interaction, and electromagnetism. ... Oh wrong theory.
Nevertheless adding solar watts to the whole mix will be interesting to determine whole system efficiency.
Wolf